NAME
    Bio::Sampling::Valection

VERSION
    version 1.0.0

DESCRIPTION
    Bio::Sampling::Valection contains a variety of algorithms for
choosing verification candidates ('calls') from competing tools or
parameterizations ('callers'), to fairly assess their performance against
each other. Originally created to select from single nucleotide variant
(SNV) calls generated by SNV mutation calling algorithms, this software
can easily be extended to other mutation types (e.g. structural variants,
gene fusions, etc.) provided data is formatted correctly. Six sampling
methods are made available to the user:

    1) Random Sampling: each call has an equal probability of getting
       selected.

    2) Equal-per-caller: an equal number of calls is selected from each
       caller.

    3) Equal-per-overlap: calls are grouped the number of callers that made
       the call, then an equal number of calls is selected from each group.

    4) Increasing-with-overlap: the likelihood of a call getting selected is
       proportional to the number of callers that made the call.

    5) Decreasing-with-overlap: the likelihood of a call getting selected is
       inversely proportional to the number of callers that made the call.

    6) Directed sampling: a) an equal number of calls is selected from each
       caller and b) the likelihood of a call getting selected is proportional
       to the number of callers that made it.

INSTALLATION
    This package requires the valection C library which can be downloaded from <http://labs.oicr.on.ca/boutros-lab/software/valection>. To install the C library on Linux, run:

    	./configure
    	make
    	sudo make install
    	sudo ldconfig

    This will install the program as 'valection', as well as the libvalec library.

AUTHOR
    Chris Cooper "<chris.cooper@oicr.on.ca>"

CONTACT
    Paul C. Boutros "<paul.boutros@oicr.on.ca>"

ACKNOWLEDGEMENTS
    Paul C. Boutros "<paul.boutros@oicr.on.ca>"
    Ontario Institute for Cancer Research

COPYRIGHT
    Copyright (c) 2015-2017 Ontario Institute for Cancer Research.  All rights reserved.